BACKGROUND: Qualitative assessments of standing plain radiographs are frequently used to determine treatment strategies and assess outcomes for the management of a wide range of foot and ankle conditions in children. A quantitative technique for such analyses would presumably be more precise and reliable. The goal of this study was to compare qualitative and quantitative techniques for the assessment of plain radiographs of the foot and ankle in children with hemiplegic type cerebral palsy (CP). METHODS: Standing anteroposterior and lateral radiographs of the foot and ankle of the involved side for 49 children with hemiplegic CP were analyzed qualitatively by 2 pediatric orthopaedists, based upon a 3-segment (hindfoot, midfoot, and forefoot) foot model. Quantitative assessment of the same radiographs was performed by 2 examiners, using 6 radiographic measurements developed to describe the alignment of the foot based upon the same 3-segment model. Intraobserver and interobserver reliability was determined for both the qualitative and the quantitative techniques. The qualitative and quantitative techniques were compared to determine agreement. RESULTS: The qualitative technique demonstrated poor-to-fair interobserver reliability (percent agreement range, 23%-31%; weighted kappa range, 0.291-0.568). The quantitative technique demonstrated good-to-excellent intraobserver (correlation coefficient range, 0.81-0.99) and interobserver (correlation coefficient range, 0.81-0.97) reliability. Percent agreement between the quantitative and the qualitative techniques for the assessment of foot segmental alignment for each examiner ranged from 22.2% to 100% (mean agreement for examiner 1 was 51% [correlation coefficient range, 0.04-0.48]; mean agreement for examiner 2 was 65.3% [correlation coefficient range, 0.22-0.85]). Percent agreement between the quantitative technique and both observers ranged from 11.1% to 83.3% (mean agreement was 36.7% [correlation coefficient range, 0.17-0.94]). DISCUSSION: Reliable quantitative radiographic analysis of the segmental alignment of the involved foot and ankle in children with hemiplegic CP is possible and is more precise and reliable than traditional qualitative techniques. Quantitative techniques can identify a wider range of foot segmental malalignments and should facilitate deformity analysis, preoperative planning, assessment of outcome, and comparison of results between centers.
BACKGROUND: Qualitative assessments of standing plain radiographs are frequently used to determine treatment strategies and assess outcomes for the management of a wide range of foot and ankle conditions in children. A quantitative technique for such analyses would presumably be more precise and reliable. The goal of this study was to compare qualitative and quantitative techniques for the assessment of plain radiographs of the foot and ankle in children with hemiplegic type cerebral palsy (CP). METHODS: Standing anteroposterior and lateral radiographs of the foot and ankle of the involved side for 49 children with hemiplegic CP were analyzed qualitatively by 2 pediatric orthopaedists, based upon a 3-segment (hindfoot, midfoot, and forefoot) foot model. Quantitative assessment of the same radiographs was performed by 2 examiners, using 6 radiographic measurements developed to describe the alignment of the foot based upon the same 3-segment model. Intraobserver and interobserver reliability was determined for both the qualitative and the quantitative techniques. The qualitative and quantitative techniques were compared to determine agreement. RESULTS: The qualitative technique demonstrated poor-to-fair interobserver reliability (percent agreement range, 23%-31%; weighted kappa range, 0.291-0.568). The quantitative technique demonstrated good-to-excellent intraobserver (correlation coefficient range, 0.81-0.99) and interobserver (correlation coefficient range, 0.81-0.97) reliability. Percent agreement between the quantitative and the qualitative techniques for the assessment of foot segmental alignment for each examiner ranged from 22.2% to 100% (mean agreement for examiner 1 was 51% [correlation coefficient range, 0.04-0.48]; mean agreement for examiner 2 was 65.3% [correlation coefficient range, 0.22-0.85]). Percent agreement between the quantitative technique and both observers ranged from 11.1% to 83.3% (mean agreement was 36.7% [correlation coefficient range, 0.17-0.94]). DISCUSSION: Reliable quantitative radiographic analysis of the segmental alignment of the involved foot and ankle in children with hemiplegic CP is possible and is more precise and reliable than traditional qualitative techniques. Quantitative techniques can identify a wider range of foot segmental malalignments and should facilitate deformity analysis, preoperative planning, assessment of outcome, and comparison of results between centers.